Monopolar and Bipolar Membranes in Organic Bioelectronic Devices

Monopolar and

Bipolar Membranes in

Organic Bioelectronic Devices

Erik O. Gabrielsson

Monopolar and Bipolar Membranes in Organic Bioelectronic Devices Erik O. Gabrielsson

During the course of the research underlying this thesis, Erik O. Gabrielsson was enrolled in Forum Scientium, a multidisciplinary doctoral program at Linköping University, Sweden.

Linköping Studies in Science and Technology. Dissertations. No. 1620 ©2014 Erik O. Gabrielsson unless otherwise noted

Typeset by the author using LA_TEX Cover by Erik O. Gabrielsson

Printed by LiU-Tryck, Linköping, Sweden, 2014 ISBN: 978-91-7519-244-4

ISSN: 0345-7524

It never gets easier, you just go faster.

Abstract

In the 1970s it was discovered that organic polymers, a class of materials otherwise best know as insulating plastics, could be made electronically conductive. As an alternative to silicon semiconductors, organic polymers offer many novel features, characteristics, and opportunities, such as producing electronics at low costs using printing techniques, using organic chemistry to tune optical and electronic properties, and mechanical flexibility. The conducting organic polymers have been used in a vast array of devices, exemplified by organic transistors, light-emitting diodes, and solar cells. Due to their softness, biocompatibility, and combined electronic and ionic transport, organic electronic materials are also well suited as the active material in bioelectronic applications, a scientific and engineering area in which electronics interface with biology. The coupling of ions and electrons is especially interesting, as ions serve as signal carriers in all living organisms, thus offering a direct translation of electronic and ionic signals. To further enable complex control of ionic fluxes, organic electronic materials can be integrated with various ionic components, such as ion-conducting diodes and transistors.

This thesis reports a background to the field of organic bioelectronic and ionic devices, and also presents the integration of ionic functions into organic bioelectronic devices. First, an electrophoretic drug delivery device is presented, capable of delivering ions at high spatiotemporal resolution. The device, called the organic electronic ion pump, is used to electronically control amyloid-like aggregation kinetics and morphology of peptides, and offers an interesting method for studying amyloids in vitro. Second, various ion-conducting diodes based on bipolar membranes are described. These diodes show high rectification ratio, i.e. conduct ions better for positive than for negative applied voltage. Simple ion diode based circuits, such as an AND gate and a full-wave rectifier, are also reported. The AND gate is intended as an addressable pH pixel to regulate for example amyloid aggregation, while the full-wave rectifier decouples the electrochemical capacity of an electrode from the amount of ionic charge it can generate. Third, an ion transistor, also based on bipolar membranes, is presented. This transistor can amplify and control ionic currents, and is suitable for building complex ionic logic circuits. Together, these results provide a basic toolbox of ionic components that is suitable for building more complex and/or implantable organic bioelectronic devices.

Populärvetenskaplig Sammanfattning

Italienaren Luigi Galvani experimenterade under 1700-talet med vad han kallade animalisk elektricitet, eller bioelektricitet. Genom att vidröra en död grodas benmuskler med olika metallelektroder kunde Galvani framkalla rörelser i grodans ben. Detta kan anses vara starten för användandet av elektricitet för att studera, mäta eller stimulera levande biologiska system och processer. Det senaste århundradets tekniska utveckling har gett oss flera medicinska applikationer där elektroder används, exempelvis i pacemakers för att reglera hjärtmusklernas sammandragning.

För vissa biologiska användningsområden är dock inte metallelektroder ideala. Detta gäller för implantat av elektroder i hjärnan då det kan bildas ärr runt elektroden som både skadar vävnaden och försämrar elektrodens förmåga att stimulera eller mäta. Konjugerade polymerer är en annan typ av elektrodmaterial som har visat sig vara användbar i biologiska applika-tioner. Dessa är plastmaterial uppbyggda av kolkemi, och är därför relativt lika biologiskt levande material. Till skillnad från isolerande plaster kan konjugerade polymerer göras elektriskt ledande. Detta upptäcktes av Alan J. Heeger, Alan G. MacDiarmid och Hideki Shirakawa på 1970-talet och belönades med Nobelpriset i kemi år 2000. De elektriskt ledande konjugerade polymererna används idag i bland annat transistorer, lysdioder och solceller. De är mjukare än metaller och är därför mer mekaniskt kompatibla med exempelvis (den ännu mjukare) hjärnan, och kan implanteras och användas under lång tid utan att ärrbildning sker.

I konjugerade polymersystem kan elektriska och joniska laddningar inte-greras och ge kombinerad elektrisk och joniska ledningsförmåga. Små och stora joner finns överallt i levande organismer, och den joniska lednings-förmågan i polymerer ger oss därför en möjlighet att översätta biologiska joniska signaler till elektroniska, och vice versa. En annan typ av jonle-dande material är jonselektiva membran. Dessa är också polymerer, och innehåller fasta orörliga positiva eller negativa joniska laddningar. De fasta laddningarna attraherar joner av motsatt polaritet, medan joner med samma polaritet repelleras. Därmed är det främst laddningar av motsatt polaritet som transporteras i det jonselektiva membranet. Om två membran med mot-satt jonselektivitet kombineras, fås ett bipolärt membran som kan användas för att bygga joniska dioder och transistorer.

I denna avhandling kombineras konjugerande polymerer med jonselektiva membran för att ge ökad kontroll över jontransport. Först visas hur enkla jonselektiva membran kan användas för att leda en ström av joner till ett mikrometerstort utlopp, varifrån jonerna sedan diffunderar ut. Den lokala ökningen i jonkoncentration används för att elektroniskt kontrollera bildningen av amyloid-liknande proteinaggregat. Amyloida plackstrukturer

återfinns i flera sjukdomar, främst då i åldersrelaterade demenssjukdomar som Alzheimers och Parkinsons sjukdom. Möjligheten att kunna kontrollera hur och var aggregeringen sker kommer förhoppningsvis att vara ett användbart redskap i sökandet efter orsaken till varför amyloider bildas och hur det kan förhindras.

Vidare beskrivs jondioder och jontransistorer baserade på bipolära mem-bran. Dessa fungerar likt sina elektroniska förlagor, men kan likrikta re-spektive modulera joniska strömmar. Jondioderna används för att skapa joniska kretsar; en AND-grind och en helvågslikriktare. AND-grinden an-vänds som en adresserbar pixel för att skapa pH-signaler, och är exempelvis lämplig för att inducera amyloidaggregering av proteiner. Helvågslikriktaren löser ett vanligt förekommande problem i applikationer där kontinuerliga jonströmmar behövs under längre tid. För att åstadkomma dessa strömmar används ofta elektrokemiska reaktioner vid metallelektroder, varpå oönskade sidoreaktioner ger upphov till gasbildning, pH-avvikelser eller giftiga sido-produkter. Genom att istället periodiskt växla strömriktnignen hos ett par konjugerade polymerelektroder och sedan likrikta strömmen med den joniska helvågslikriktaren undkommer man dessa problem. Därmed kan generering av en konstant jonström frikopplas från den elektrokemiska kapaciteten hos elektroderna som används.

Tillsammans ger dessa resultat en verktygslåda av jonledande kompo-nenter som lämpar sig för att öka möjligheten att joniskt kommunicera med levande organismer, exempelvis i framtida implanterbara biomedicinska komponenter.

Acknowledgements

This thesis would not have been possible without help and support from people around me. For this I like to express my sincere gratitude to: Magnus Berggren, my main supervisor, for giving me the opportunity to work in the Laboratory of Organic Electronics and to evolve as a scientist, and for your optimism and great ideas.

Edwin Jager and Daniel Simon, my first and second co-supervisors, for their support and for sharing their knowledge in bioelectronics.

Peter Nilsson and Per Hammarström, for their invaluable knowledge in amyloids, their enthusiasm, and experimental help.

Sophie Lindesvik, for being such a great and friendly administrator. Lars Gustavsson, Bengt Råsander, Anna Malmström, and all the other personal that ensures that the cleanroom remain tidy and up-and-running.

All present and former members of the Laboratory of Organic Electron-ics for their part in creating an amazing and friendly research environment. Especially, I would like to thank: Klas, “Dr. Iontronic”, for his expertise in ion transport, and for all the fun times (not only in the lab). Pelle, for being such a truly nice guy with awesome pipetting skills. Malti, for the fun friday experiments and coffee table discussions. Anders Hentzell for the Friday lunches. Maria and Kristin, for the great times on conferences and study visits. Astrid, for the enthusiastic help with the Aβ experiments. Amanda and Theresia for their contribution in development of new pro-cesses and materials. Xiaodong and Skomantas for sharing office with me, and making sure I never wanted my own room.

The staff of Acreo in Norrköping, especially Anurak Sawatdee for the hugs and company during commutes, David Nilsson for his expertise in inkjet printing, and Mats Sandberg for discussions and help regarding everything related to chemistry.

Annelie Eveborn and Olle-Jonny Hagel at Thin Film Electronics for advise on cleanroom processing.

Forum Scientum and Stefan Klintström, for arranging both scientifically valuable and fun activities.

The cycling team CK Hymer and all my other cycling friends, for the amazing and fun hours training and competing, in sunshine, rain, headwinds, and snow.

All my friends, for their friendship and fun times. Special thanks to Rebecka, Stina, Amie, and Gustav for making Linköping such a nice town.

My family; my sister, my brothers, and my parents Kerstin and Lennart, for all love and support.

Annika, for love and being loved and for understanding me, and to our wonderful “kleiner Mann” Wilhelm for bringing so much joy.

Tack!

Included Papers

Paper I

Spatially Controlled Amyloid Reactions Using Organic Electronics Erik O. Gabrielsson, Klas Tybrandt, Per Hammarström, Magnus Berggren, and K. Peter R. Nilsson

Small 6, pp. 2153–2161 (2010)

Contribution: Contributed to the experiment design, performed most experi-mental work, wrote the first draft, and contributed to the editing of the final manuscript.

Paper II

Controlled Microscopic Formation of Amyloid-Like Aβ Aggregates Using an Organic Electronic Device

Erik O. Gabrielsson, Astrid Armgarth, Per Hammarström, K. Peter R. Nilsson, and Magnus Berggren

Manuscript in preparation

Contribution: Designed experiments, performed and supervised parts of the experimental work, and wrote the first draft.

Paper III

Ion diode logics for pH control

Erik O. Gabrielsson, Klas Tybrandt, and Magnus Berggren

Lab on a Chip 12, pp. 2507–2513 (2012)

Designed experiments, performed all experimental work, wrote first draft of manuscript, and contributed to the editing of the final manuscript.

Paper IV

Polyphosphonium-Based Bipolar Membranes for Rectification of Ionic Currents

Erik O. Gabrielsson and Magnus Berggren

Biomicrofluidics 7, 064117 (2013)

Contribution: Designed experiments, performed all experimental work, and wrote most of the manuscript.

Paper V

A Four-Diode Full-Wave Ionic Current Rectifier Based on Bipolar Membranes: Overcoming the Limit of Electrode Capacity

Erik O. Gabrielsson, Per Janson, Klas Tybrandt, Daniel. T. Simon, and Magnus Berggren

Advanced Materials 26, pp. 5143–5147 (2014)

Contribution: Designed experiments, performed most experimental work, and wrote most of the manuscript.

Paper VI

Polyphosphonium-Based Ion Bipolar Junction Transistors Erik O. Gabrielsson, Klas Tybrandt, and Magnus Berggren Submitted manuscript

Contribution: Designed experiments, performed most experimental work, and wrote most of the manuscript.

Related Work

Toward Complementary Ionic Circuits: The npn Ion Bipolar Junction Transistor

Klas Tybrandt, Erik O. Gabrielsson, and Magnus Berggren

Journal of the American Chemical Society 133, pp. 10141–10145 (2011)

Ultra-low voltage air-stable polyelectrolyte gated n-type organic thin film transistors

Abdellah Malti, Erik O. Gabrielsson, Magnus Berggren, and Xavier Crispin

Contents

I

Background

1 Introduction 1

1.1 Ions and Organic Bioelectronics . . . 1

1.2 Aim of Thesis . . . 1 1.3 Outline of Thesis . . . 2 2 Conjugated Polymers 3 2.1 Orbitals . . . 3 2.1.1 Atomic Orbitals . . . 3 2.1.2 Molecular Orbitals . . . 4 2.1.3 Hybridization . . . 4 2.2 Conjugated Polymers . . . 5 2.2.1 Electronic Structure . . . 5

2.2.2 Charge Carriers and Transport . . . 7

2.2.3 Doping . . . 8

3 Ion Transport 11 3.1 Ion Transport Processes . . . 11

3.1.1 Migration and Diffusion . . . 12

3.1.2 Transport Numbers . . . 12

3.2 Ion Selective Membranes . . . 12

3.2.1 Donnan Potential . . . 14

3.2.2 Junction Potential . . . 15

3.2.3 Membrane Potential . . . 15

3.2.4 Concentration Polarization . . . 15

3.3 Bipolar Membranes . . . 16

3.3.1 Forward Bias Regime . . . 17

3.3.2 Reverse Bias Regime . . . 18

3.3.3 Rectification . . . 18

3.3.4 Electric Field Enhanced Water Dissociation . . . 18

3.4 Electrodes in an Electrolyte . . . 20

4 Amyloids 23 4.1 Amyloid Composition and Structure . . . 23

4.2 Fibrillogenesis . . . 25

4.3 Factors for Amyloid Formation . . . 27

4.4 Pathogenesis . . . 27

4.4.2 Non-Neurophatic and Systematic Amyloid Disease . . 28

4.4.3 Functional Amyloids . . . 28

4.5 Detection Methods . . . 28

4.6 Amyloids as Nano-Structural and Functionalized Material . . 30

4.7 Lab-on-a-Chip Technology in Amyloid Research . . . 30

5 Organic Bioelectronics 31 5.1 Sensors . . . 31 5.2 Actuators . . . 33 5.2.1 Electronic Stimulation . . . 33 5.2.2 Mechanical Actuators . . . 33 5.2.3 Surface Switches . . . 34

5.2.4 Drug Delivery Systems . . . 35

6 Experimental Methods 37 6.1 Microfabrication Techniques . . . 37 6.1.1 Spin coating . . . 37 6.1.2 Photolithography . . . 38 6.1.3 Dry Etching . . . 38 6.1.4 Inkjet Printing . . . 40

6.2 Synthesis of Anion Exchange Membranes . . . 40

6.3 Fabrication of Ion Pumps, Diodes, and Transistors . . . 41

7 Devices in Papers 45 7.1 General Design . . . 45

7.2 Organic Electronic Ion Pumps . . . 45

7.2.1 In Vitro Stimulation of Cells . . . 46

7.2.2 In Vivo Stimulation of Tissues . . . 46

7.2.3 Formation of Amyloid-Like Aggregates (Papers I-II) . 47 7.3 Ion Diodes . . . 48

7.3.1 The IBMDs (Papers III-IV) . . . 48

7.3.2 Alternatives to IBMDs . . . 51

7.4 Ion Diode AND Gate (Paper III) . . . 52

7.5 Ion Diode Full-Wave Rectifier (Paper V) . . . 54

7.6 Ion Bipolar Junction Transistors (Paper VI) . . . 55

8 Concluding Remarks 59 8.1 From Amyloids to Full-Wave Rectification . . . 59

8.2 Major Findings . . . 60 8.3 Future Outlook . . . 61 References 63 Acronyms 75

II

Papers

Part I

Chapter 1

Introduction

1.1 Ions and Organic Bioelectronics

In 1771 Luigi Galvani discovered what he called “animal electricity”. In the associated experiments he applied metal electrodes to detached frog legs and caused twitching movements in the legs. Galvani’s work laid the ground for studies of electronic stimulation of biological system, and conversely, also the recording of neuronal signals. Centuries later, electrodes are, for example, used for stimulation of heart rhythm in pacemakers, combined with biological enzymes to provide glucose sensing for diabetic patients, and for recording of membrane potential changes in signaling neuronal cells. The research field in which electronic devices is used to record and regulate biological processes is today know as bioelectronics.

Traditional materials for electronic devices are metals and silicon, which provide high conductivity and can be patterned at high resolution. Unfor-tunately, metal and silicon have some important limitations with respect to the integration with biology. For example, metals and silicon are hard materials, while most biological tissues are soft. A possible solution to the hard/soft-incompatibility problem was found in the 1970’s when Alan J. Heeger, Alan G. MacDiarmid, and Hideki Shirakawa discovered that poly-acetylene could be rendered electronically conductive through treatment with iodine; a discovery that was awarded the Nobel price in chemistry in 2000. As polyacetylene is a carbon-based polymeric material the new research field of conductive polymers was called organic electronics. Today, the field of organic electronics has produced numerous technologies and devices, such as organic light emitting diodes, organic field effect transistors, organic solar cells, and organic memories.

The polymeric materials used in the field of organic electronics have some unique properties that are not typically found in regular silicon based elec-tronics, such as solution processing, softness and flexibility, and a combined electronic and ionic conductivity. Their softness, carbon-based chemistry, and possibility for chemical alterations enable the organic electronic materials to be highly biocompatible. Further, the importance of ions, as signals in biological systems, make the ion- and electron-conducting polymers a perfect translator of signals in bioelectronic stimulation and sensing applications. The fruitful combination of organic electronics in bioelectronic applications became known as organic bioelectronics.

1.2 Aim of Thesis

This thesis aims to explore the combination of ion-conducting membranes and organic electronics to allow for greater control of ionic currents in

bioelectronic devices. The possibility to direct a flux of ions into a spatially well-defined release point is presented as a novel method to study ion-induced aggregation of amyloid proteins. Additionally, ion conductive membranes with opposite polarity are used to realize devices with function similar to electronic semiconductor diodes and transistors, but for ionic currents and signals. These non-linear ionic devices are further investigated for constructing chemical circuits where the current within the circuit is entirely ionic to its nature.

1.3 Outline of Thesis

The first part of this thesis describes the background to the field of organic bioelectronics, with a focus on ionic components and applications within the field of amyloid research, followed by a more in-depth description and explanation of the function of novel bipolar membrane (BM) based ionic devices. In the next chapter, Chapter 2, the concept of conjugated polymers (CPs), and why they can be electronically conductive, is explained. Chapter 3 aims at presenting a similar theoretical background to ion conductivity in electrolytes, with emphasis on ion exchange membranes (IEMs). Chapter 4 provides a short introduction to the field of amyloids, and Chapter 5 gives a brief survey over the field of organic bioelectronics. The experimental techniques used in this work are described in Chapter 6. Chapter 7 presents the background and function of the ionic devices used in the included papers, giving a detailed view of the work produced for this thesis. The last chapter, Chapter 8 serves to provide my own reflections of the presented work, with some analysis of its impact and thoughts about future development of ionic devices as well as their potential applications.

The second part of the thesis includes the manuscripts written and pub-lished during the work towards this thesis. Papers I–II describe the use of the organic electronic ion pump to produce amyloid-like protein aggregates with spatiotemporal control. Papers III–V present the development, challenges, and improvements of ion diodes based on BMs, as well as the integration of said diodes into simple ionic circuits (AND logic gate in Paper III, full-wave rectifier in Paper V). Paper VI details the development of ion transistors, also based on BMs, where the transistor performance is improved by reducing the dimensions of the BM.

Chapter 2

Conjugated Polymers

Polymers are macromolecules consisting of several (poly) repetitions of small subunits (mers). Many polymers are materials that we encounter in our everyday life and typically describe as plastic insulating materials. Some polymers however, namely CPs, can be electronically conducting. The difference between an insulating polymer and a CP is in the electronic structures, i.e. the nature of the molecular bonds between the atoms in the polymers. This chapter describes how these bonds are formed and how CPs can be made electronically conducting.

2.1 Orbitals

2.1.1 Atomic Orbitals

An atom contains positively charged protons, negatively charged electrons, and (most often) neutral neutrons. The number of protons and electrons are equal, so that the total charge of the atom is zero. Protons and neutrons are located at the center of the atom, forming the nucleus, and electrons form a cloud surrounding the nucleus. By describing the atom using quantum mechanics, discrete states of the electrons are obtained [1]. These are arranged into atomic orbitals (AOs), defined by shell and orbital type. The shells (K, L, M, N, etc. or 1, 2, 3, 4, etc.) describe the distance of the orbital from the nucleus, where K or 1 is the innermost shell. Each shell can contain a number of orbital types (s, p, d, etc.) with different spatial distribution.

Quantum mechanics dictates that each AO can only be occupied by two electrons and only if they have different spin (up or down). Further, the energy for an electron in an AO is typically increased with shell number and orbital type (e.g. 2s have higher energy than 1s, 2p have higher energy than 2s). Therefore, electrons fill the innermost shells first to minimize the energy of the atom, resulting in the ground state of the atom. For a carbon atom, with 6 electrons, this gives a ground state configuration of 1s2 _2s2 2p2 _{(Figure 2.1). The superscript denotes the number of electrons in each} orbital. Electrons located in the outermost shell, the valence shell, are called valence electrons. The chemistry of atoms and molecules is highly dictated

x y z

2s 2px 2py 2z

1s

1s

1˜

1˜*

Ener

gy

Figure 2.2: Energy diagram for two 1s orbitals (e.g. two hydrogen atoms) forming a

1σ MO and a 1σ* MO. The electrons from the 1s orbitals occupy the 1σ MO, forming a bond (e.g. a hydrogen molecule). Constructive interference between the AOs results in lower MO energy (1σ), while destructive interferance results in higher MO energy (1σ*).

by interactions between valence electrons. In most organic molecules the valence orbitals are formed from s and p orbitals.

2.1.2 Molecular Orbitals

Molecules are formed by atoms bonding to each other. When two atoms are brought close to each other their valence orbitals start to overlap and the valence electrons can interact. The interaction can be approximated by a linear combination of the individual AOs, and the result is formation of molecular orbitals (MOs) (Figure 2.2) [2]. MOs are shared across the participating atoms, and are populated by the electrons from the contributing AOs. Depending on the involved AOs different MOs are formed. A σ orbital is formed if the bonding AOs are arranged along the axis of the bond, and a π orbital if the arrangement is perpendicular to the bond axis. The σ and π orbitals have spatial similarities to s and p orbitals, respectively.

The interaction between AOs results in MOs of different energy as the energy levels split up into bonding and antibonding orbitals. A bonding MO has lower energy than the individual AOs, and stabilizes the interaction between atoms to form a bond. An antibonding orbital, denoted with an asterisk (*) has, on the other hand, higher energy and destabilizes the bond. Thus, a molecular bond is only stable if the energy of the electrons in the bonding and antibonding MOs is lower than in the original AOs. The occupied MO with highest energy is called highest occupied molecular orbital (HOMO). The orbital directly above the HOMO in energy is the lowest unoccupied molecular orbital (LUMO).

2.1.3 Hybridization

According to valence bond theory [3] some atoms form hybridized orbitals when forming bonds, for example carbon. Here, one s and one, two, or all three of the p orbitals of carbon (Figure 2.1) can form hybridized sp, sp2_, or sp3_{orbitals (Figure 2.3). The hybridized orbitals can form σ bonds with} other atoms, while the non-hybridized p orbitals (for sp or sp2 _{hybridization)} 4

sp sp2 sp3

Figure 2.3: Illustrations of hybridized orbitals (shaded) formed by carbon. The

orbitals with dotted outline are non-hybridized p orbitals.

2p

2˜

2˜*

Ener

gy

2p

Figure 2.4: Energy diagram of two p orbitals forming π MOs.

can form π orbitals. The formed π orbitals are either bonding or antibonding (Figure 2.4).

Each occupied MO between two atoms represents a bond. More than one bond can be formed between two atoms in a molecule, for example by occupation of both σ and π MOs, giving a double bond. In ethane, all bonds are single bonds (Figure 2.5a) formed by σ MOs, while in ethylene an additional π MO between the carbons results in a double bond (Figure 2.5b).

2.2 Conjugated Polymers

2.2.1 Electronic Structure

The difference between an insulating polymer, such as polyethylene ure 2.6a), and an electronic conductive CP, such as polyacetylene

(Fig-˜

˜

°

Figure 2.5: The molecule (a) ethane consists of only single bonds, while (b) ethylene

(a) Polyethylene (C2H4)nH2 (b) Polyacetylene (C2H2)n

Figure 2.6: Structures of (a) the insulating polyethylene, containing only single

bonds, and (b) the conductive polyacetylene, containing alternating single and

double bonds.

ure 2.6b), is the hybridization of the carbon atoms. In polyethylene, the carbons are sp3 hybridized, and all valence electrons are involved in forming σ bond, either to other carbons or to hydrogens. In polyacetylene the carbons are sp2 _{hybridized. This allows for three σ bonds, two to other carbons and} one to hydrogen. The remaining p orbital is free to form a π bond with a neighboring carbon atom with an unoccupied p orbital. Thus, in polyacety-lene an alternating pattern of double bonds is formed, called a conjugated structure. π orbitals across a polymer can overlap to give further electronic interaction (Figure 2.4), resulting in additional bonding and antibonding orbitals across the polymer. As more π orbitals interact the orbital energy levels split up. The number of available levels depends on the number of carbon atoms in the polymer (Figure 2.7). As the number of carbon atoms in the backbone of a polymer increases, the energy levels of the orbitals form bands.

If the lengths of single and double bonds were equal in a CP chain, bonding and antibonding energy levels would eventually form one continuous energy band and give metallic conductivity to the polymer. However, as described by Peirel’s theorem, bond lengths are not equal between single and double bonds. Alteration of bond lengths in a CP minimizes the energy,

Ener gy

˜-band

˜*-band

CH

C

H

C

H

C

H

E

Figure 2.7: Energy diagram for the formation of π orbitals i polyacetylene of

vari-ous lenght. The number of possible combinations of constructuve and destructive interference increase with length, thus splitting the energy levels.

and is thus the preferred arrangement. This has a stabilizing effect on the bonding orbitals while destabilizing the antibonding orbitals. The energy levels of bonding and antibonding orbitals therefore separates into two bands, the π and the π* band. The HOMO is then located at the top of the π band, while the LUMO is on the bottom of the π* band. The difference in energy between HOMO and LUMO is the band gap. A small band gap is representative for a semiconducting material, and for CPs the band gap is typically 1.5–3 eV [4]. For an insulating material, such as polyethylene, the band gap is larger.

2.2.2 Charge Carriers and Transport

Electronic conduction requires charge carrierss to be present in the material. Depending on the structure of the polymer chain, the charge carriers in a CP are either solitons or polarons [5].

Solitons

If the location of single and double bonds in a CP can be interchanged without altering the system energy, the polymer is said to have a degenerate ground state (Figure 2.8). As both configurations are equal in energy they also have equal probability of occurring. Both configurations can even exist on the same chain. This gives a transition region between the two configurations with modified bonds, a soliton. Solitons give two energy states inside the band gap and can have either neutral, positive, or negative charge.

A B

B A

Figure 2.8: The A and the B bond configuration of polyacetylene have equal

proba-bility to occur. If both exist on the same chain a soliton is formed.

Polarons

Most CPs do not have a degenerate ground level, i.e. interchanging single and double bonds alters the polymer’s energy. An example for which this is true is the polythiophenes. The ground state of a polythiophene is the aromatic configuration where the double bonds are located in aromatic rings (Figure 2.9a). By interchanging single and double bonds, the higher energy

quinoid configuration is obtained (Figure 2.9b).

Solitons are not observed in these polymers as alternation of bond gives different energy. However, by introducing a charge to the polymer, a local deformation of the bond conjugation is obtained. This is called a polaron (Figure 2.10), and can be regarded as a local quinoid configuration, i.e.

alteration of single and double bonds around a limited number of bonds. If several polarons exist on the same chain it is sometimes energetically

S S S (a) Aromatic S S S (b) Quinoid

Figure 2.9: The (a) aromatic form of a polythiophene, where the double bonds are

located inside the ring, has lower energy than the(b) quinoid form.

favorable to locate them close together to form bipolarons. Polarons are either positive or negative depending on the charge added to the CP. The energy levels of the polarons lie between the π bands, giving the doped polymer chain a reduced band gap.(Figure 2.11).

Charge carrier transport

Solitons and polarons can with ease move along the chain of a polymer, but this motion is naturally restricted to the chain length of the polymer. To achieve charge transport over longer distances charge carriers also need to move between polymer chains. This motion is typically characterized by thermally activated hopping of the charge carriers between chains, and this process limits the electronic mobility of a CP. The energy requirement can be reduced by increasing the π overlap between chains, such as ob-tained for highly crystalline films. Because of the difference in π overlap charge mobility can vary from 10 × 10−5 to 10 × 10−4cm2/(V s) for amor-phous poly(3-hexylthiophene) [6] to 0.2–0.6 cm2/(V s) for a single-crystal polythiophene-derivate with side groups designed for promoting crystalliza-tion [7].

2.2.3 Doping

The relative large band gap of an intrinsic (undoped) CPs results in few ther-mally excited charge carriers and limited conductivity. Higher conductivity can be obtained by introducing more charge carriers into the polymer chain in a process called doping [8]. Doping involves electron transfer between a dopant species and the polymer. If electrons are transferred from the dopant to the polymer, i.e. reduction of the polymer, n-doping occurs. Conversely, transferring electron from the polymer to the dopant, i.e. oxidation of the polymer, is called p-doping. The population of charge carriers, i.e. polarons and/or solitons, increases with doping level. The introduced charge in the polymer is compensated by an ion with opposite charge.

S S S S S S + Positive polaron

Figure 2.10: A local quinoid configuration gives a (positive) polaron on the chain.

˜-band ˜*-band

Positive polaron

Neutral _bipolaronPositive Negative_{polaron bipolaron}Negative

Figure 2.11: Energy diagrams of positive and negative polarons and bipolarons.

In chemical doping, the dopant molecule has a HOMO/LUMO higher/ lower than the LUMO/HOMO of the polymer, and is therefore oxidized/ reduced. The oxidized/reduced dopant becomes the compensating positive/ negative ion to the charge introduced in the CP chain. The halogen doped polyacetylene studied by Shirakawa et al. [9] is an example of a chemically doped conducting polymer.

Electrochemical doping is performed by submerging a CP electrode in an electrolyte [10]. By applying a voltage between the CP electrode and a working electrode, electric charge is introduced into the CP through electro-chemical oxidation or reduction. The introduced charges are compensated by ions from the electrolyte.

For a partially doped CP electrode only a small voltage change (vs. a counter electrode) is needed to change the oxidation/reduction state of the polymer electrode, i.e. the doping level can be changed with ease. This is advantageous for numerous applications, among them bioelectronic de-vices [11], as a small change in potential at the electrode interface can lead to an observable change in electronic properties of the CP electrode.

Chapter 3

Ion Transport

An electrolyte is a solution containing dissolved ions, i.e. positive cations (e.g K+) and negative anions (e.g Cl–). Each ion can have one or multiple charges, but the number of positive and negative charges in the electrolyte must to be equal to maintain electroneutrality. The ions can be simple metal ions (Figure 3.1a), more complex organic molecules such as charged biomolecules, or a polymer chain containing repetitive charges (as in a polyelectrolyte, Figure 3.1b). A binary electrolyte contains two ionic species in the same amount, e.g. KCl. A ternary electrolyte contains three ionic species, e.g. a solution of NaCl mixed with KCl or a polyelectrolyte in NaCl. The conductivity of an electrolyte increases with the ion concentration as long as the ions do not form ion-pairs [12].

+ – + + – – + – – +

(a) A simple electrolyte

+ – + + – – + – + – (b) A polyelectrolyte

Figure 3.1: Examples of electrolytes. The black lines represents polymer chains,

onto which (positive) charges are fixated

3.1 Ion Transport Processes

In an electrolyte, ions can move by three processes; migration, diffusion, and convection [13]. Migration occurs in presence of an electric field (Figure 3.2a), net transport by diffusion occurs along concentration gradients (Figure 3.2b), and convection occurs due to fluid movement (Figure 3.2c).

+ – + + – –

+

–

3.1.1 Migration and Diffusion

In a solution with a concentration difference ∆c, random motion of molecules gives diffusion, where the net movement of molecules will strive to level any concentration difference. For example, if a volume with high electrolyte concentration is put into contact with a volume with low concentration, ions will diffuse from high to low concentration. Flick’s first law describes the diffusional flux (J ) of molecules due to a concentration gradient ∆c as:

J = −D∆c (3.1)

where D is the diffusion coefficient, a property which describes the ability of a molecule to move in its solvent.

A potential difference ∆φ in an electrolyte gives an electric field that will exert a force on the ions in the electrolyte, resulting in ion migration. For cations (z = +1) the migration is in the direction of the electric field, and vice versa for anions (z = −1).

Migration and diffusion are linked for charged species such as ions: A flux of ions due to migration creates a concentration gradient, resulting in a diffusive component for the involved ions. Similarly, diffusion of two oppositely charged ions at different speed induces an electric field in the electrolyte, giving migration. The total flux of an ion in an electrolyte is described by the Nerst-Planck equation:

J = −D∆c + zF

RTDc∆φ (3.2)

where z and c the charge and concentration of the ion, respectively, F is Faraday’s constant, R is the gas constant and T is the temperature. The first part of the equation describes diffusional flux and the second part describes migrational flux.

3.1.2 Transport Numbers

As an electrolyte is composed of two or more types of ions, the total ionic transport through the electrolyte is described by the sum of all individual fluxes. Ions in an electrolyte can have different z, D, c, or ∆c, and they therefore contribute differently to the total flux. The transport number defines the fraction of the total flux for a specific ion the electrolyte. For electrolytes containing ions with roughly equal diffusion coefficients (e.g. for KCl, D = 2.0 × 10−5cm2_{/s [14]) and no concentration gradients, the} transport numbers for both cation and anion are approximately 0.5. For a HCl-electrolyte, where the diffusion coefficient for H+ is significantly larger (D = 9.3 × 10−5cm2_{/s) the transport number of H}+

is higher than for Cl– (tH+= 0.821 vs. t_Cl− = 0.179) [15].

3.2 Ion Selective Membranes

Membranes are often used to separate electrolytes. Some membranes are non-selective, i.e. all species in the electrolytes can pass freely (Figure 3.3a), while others offer higher permeabeability to specific species [16]. An IEM, 12

+ – + – + –

+

–

(a) Neutral membrane

– –

+

–

+

–

Figure 3.3: Migrational transport through a (a) non-selective membrane, a (b) CEM

and an(c) AEM. In the IEMs, ions are blocked or let through based on their polarity.

separating two electrolytes, offer higher permeabeability for ions with certain charge to cross between the electrolytes, i.e. is either a cation exchange membrane (CEM) (Figure 3.3b) or an anion exchange membrane (AEM) (Figure 3.3c). The ion selectivity in an IEM often arises from fixed ion charges in the membrane in the form of a polyelectrolyte. Commonly used fixed charge groups are the negatively charged sulfonate group and the positively charged quaternary amine group (Figure 3.4) [17].

S O O O– R1 Na + (a) Sulfonate N+ R3 R2 R4 R1 Cl– (b) Quaternary amine

Figure 3.4: Common (a) negative and (b) positive fixed charged groups in

polyelectrolyte-based IEMs.

The non-fixed ions in the membrane are called mobile ions, and are either of same charge (co-ions) or opposite charge (counter-ions) to the fixed charges. Local electroneutrality dictates that the sum of ionic charges should be zero. The fixed charges are therefore compensated by an equal amount of counter-ions from the electrolyte. At the same time, the fixed charges repel the similarly charged co-ions. This creates an unequal distribution of co-ions and counter-ions in the membrane; there are more counter-ions than co-ions. The fraction of counter-ions in an IEM can be calculated based on the fixed

charge concentration X and external electrolyte concentration cs: ccou ccou+ cco = 1/2 + 1 4p1/4 + (cs/X)2 (3.3)

where ccouand ccoare the counter- and co-ion concentration in the membrane,

respectively [18]. For a membrane with X = 1 m and cs= 0.1 m, the

counter-ion fractcounter-ion is ∼ 99 % of the total amount of mobile counter-ions.

The concentration difference increases the counter-ion transport number and decreases the co-ion transport number as compared to a free electrolyte. As the fixed charges in the membrane cannot move, an ionic current through the membrane is carried by the mobile ions, and predominantly by the counter-ions as they have higher concentration, i.e. the membrane is selective to counter-ions. The selectivity will depend on concentration of fixed charges in the membrane, surrounding electrolyte concentration, and magnitude off the ionic current through the membrane [16]. Typical highly charged IEMs have fixed charge concentrations of 1–3 m and counter-ions account for 95–99 % [19] of the current, while for less charged membranes the selectivity drops.

3.2.1 Donnan Potential

When an IEM is in contact with an electrolyte, there will be a concentration gradient for both counter- and co-ions between the membrane and the electrolyte; the concentration of counter-ions is (typically) higher in the membrane than in the electrolyte and vice versa for co-ions [18]. This leads to counter-ions and co-ions striving to diffuse out of and into membrane, respectively, and bending the concentration profiles near the interface. The movement of ions charges the interface, giving an electric field that opposes the diffusion (Figure 3.5a). The produced potential drop over the membrane interface is called the Donnan potential and is described by:

ϕDon = RT zFln cs ccou (3.4) where z and cs_{is the charge and electrolyte concentration of the counter-ion.}

+ – Pot en tial ˜Don Conc en tr ation Distance ccou cs c co

(a) Donnan potential

+ – Conc en tr ation Pot en tial ˜Di° Distance cs1 cs2 (b) Junction potential Conc en tr ation Pot en tial ˜Di° ˜_Don1 ˜Don2 Distance cs1 cs2 ccou cco (c) Membrane potential Figure 3.5: Potentials arising across a (cation-selective) membrane.

3.2.2 Junction Potential

If a concentration gradient exists in an electrolyte (e.g. if a membrane separates two electrolytes of different concentrations) ions will diffuse from higher to lower concentrations [18]. Under zero-current conditions, i.e. when no net-charge is transported through the membrane, the flux of cations and anions need to be equal to maintain electroneutrality. Therefore, the flux of an ion with higher diffusion coefficient cannot be larger than for its counter-ion. This creates a potential difference along the concentration gradient that opposes the diffusion of the faster ion (Figure 3.5b). This potential is called the junction or diffusion potential and is described by:

ϕdif f = (t−− t+)

RT zFln

cs2

cs1 (3.5)

where t− and t+ are the transport numbers of the anion and cation in the electrolyte and cs1 _{and c}s2 _{are electrolyte concentrations. For a}

HCl-electrolyte, with t+ = 0.821 and t− = 0.179, this gives ϕdif f = −38.0 mV

per decade of cs2_/cs1 _{at 25}◦_{C, while for KCl (t}

+ = 0.4906, t− = 0.5094)

ϕdif f = +1.1 mV per decade [15].

3.2.3 Membrane Potential

In total, three potentials exist across a membrane separating two electrolytes; two junction potentials, at either side of the membrane, and one diffusion potential across the membrane (Figure 3.5c) [18]. The sum of these potentials can be measured using electrodes immersed into the electrolytes, and is called the membrane potential:

ϕm= ϕdif f+ ϕDon2− ϕ_Don1 (3.6)

If the electrolyte has similar diffusion coefficient for both anions and cations (as for KCl) the membrane potential is approximately reduced to two Donnan potentials. These are dependent on electrolyte concentrations, and can for an ideal membrane be written as:

ϕm=

RT zFln

cs1

cs2 (3.7)

The membrane potential is related to the transport number of the mem-brane [20]: ϕm= (2tcou− 1) RT zF ln cs1 cs2 (3.8)

Thus, transport numbers inside a membrane can be calculated from mea-surements of the membrane potential. For example, an ideal IEM with tcou

= 1 yields a ϕm of +17.8 mV at 25◦C when immersed between two KCl

electrolytes with cs1_/cs2 _{= 2, while a less selective membrane with t}

cou =

0.8 gives +10.7 mV.

3.2.4 Concentration Polarization

When a current is applied across an IEM, the difference in transport number for counter-ions and co-ions results in more counter-ions passing through

ccou cÿx cco cs1 cs2 Conc en tr ation I = 0 I < Ilim I = Ilim I > Ilim Distance

Figure 3.6: Concentration polarization across an IEM

the membrane than co-ions in the opposite direction. This gives rise to concentration polarization across the membrane [21]. At the feeding side of the membrane, this results in a lower ion concentration (compared to the bulk electrolyte), as counter-ions and co-ions move in opposite directions away from the membrane interface (Figure 3.6). At the opposite side, there is a corresponding increase in ion concentration. The concentration gradients become more pronounced with the applied current, i.e. lower/higher concentration at the interface and reaching farther out from the membrane. For a specific current level, the limiting current (Ilim), the concentration at the feeding interface approaches zero. For current levels above Ilimthe zero-concentration point moves farther out and introduces an additional potential loss [22]. Thus, the resistance of the IEM differs below and above Ilim.

3.3 Bipolar Membranes

A BM is constructed by layering two IEMs of opposing polarity (i.e. one CEM and one AEM, Figure 3.7a). Frilette studied this structure in 1956 by pressing two commercial oppositely charged membranes, and he observed asymmetric behavior depending on the applied voltage polarity [23]. The two oppositely charged IEMs, with fixed charge concentration cf ix− and

cf ix+, defines two regions where either cations or anions are counter-ions

and majority mobile species (c+ and c−) and the respective co-ions (c−

– + –+ + + + + + – – – – – or

(a) Sandwiched structure

c+ cÿx– c– cs1 cs2 Conc en tr ation c– cÿx+ c+ Distance (b) Concentration profile

Figure 3.7: (a) Structure of a BM, with a CEM (medium grey) and an AEM (dark grey)

region. The interface can either be well-defined or with a neutral layer inbetween.(b)

The concentration profile of a BM shows two regions with opposite mobile charge concentrations.

Voltage Current Forward bias Reverse bias EFE water dissociation

Figure 3.8: Typical current-voltage profile for a BM, with (solid line) and without

(dotted line) EFE water dissociation.

and c+, respectively) are repelled (Figure 3.7b) [24]. Thus, the majority mobile species in one region cannot easily enter the next. The area where the ion selectivity changes is called the junction of the BM. In some BMs there is a neutral region situated in the junction, separating the two IEMs (Figure 3.7a). This region can either be intentionally included or a defect

caused by the fabrication of the BM.

Unlike monopolar membranes, the mechanism for ionic transport in a BM will vary depending on the ion current’s direction, i.e. on the voltage bias applied across the BM. In general, three bias regimes are defined for BMs; forward bias, reverse bias, and electric field enhanced (EFE) water dissociation regime. These regimes give BMs a distinct current-voltage profile (Figure 3.8) [24].

3.3.1 Forward Bias Regime

In the forward bias regime, a positive voltage is applied at the CEM side of the BM vs. the AEM side. This enables mobile counter-ions on both sides of the BM to migrate towards the middle of the BM (Figure 3.9a). Additional counter-ions can enter the BM from either side of the BM. Thus, in the forward bias regime an ionic current, composed of anions and cations on either side of the BM, can flow in through the BM. The main voltage drop in the forward bias is due to the bulk resistance of the IEMs, and the current (below Ilim) is therefore expected to be linearly dependent on the applied

voltage (Figure 3.8).

Ion accumulation and co-ion crossing

Counter-ions reaching the zone where the polarity of the BM changes cannot easily continue migrating in the electric field as they then become co-ions. This causes an accumulation of ions inside the junction [25], as illustrated in Figure 3.9b. However, selectivity of an IEM is dependent on the surrounding ion concentration (Equation 3.3). As ions accumulate at the junction, co-ion transport increases and ions start to cross the BM [25] (Figure 3.9c). A large neutral region inside a BM lowers the ion concentration and delays the onset for co-ion transport.

3.3.2 Reverse Bias Regime

Reversal of the applied bias over the BM has a major impact on the ion transport process through the BM, as this reverses the migration direction of the two counter-ions in the system. The counter-ions will thus move from the junction towards either end of the BM (Figure 3.10a). As the amount of ions in the junction is finite and only a small flux of new ions can be supplied to the junction by co-ion migration, the ion concentration in the junction will decrease. Eventually, the low ion conductivity in the junction, due to the low ion concentration, limits the current through the BM. The main potential drop then occurs across the BM junction. As in the forward bias regime, the reverse bias current through the BM will approximately be linearly dependent on the applied voltage, but with a higher resistance than observed in the forward bias regime as the transport is dependent on co-ions (Figure 3.8).

Ion accumulation hysteresis

The process of changing a BM from forward bias to reverse bias regime includes depleting the junction of ions, and is thus dependent on the amount of ions stored in the junction (figures 3.10b–3.10d). For a junction where high ionic charge can be stored without inducing significant co-ion leakage, e.g. if a neutral layer is present inside the BM, the amount of ions that need to be extracted to reach the reverse bias regime vs. the amount of ions injected during forward bias can reach close to 100 % [26]. This causes a delay in the forward bias/reverse bias transition of the BM, as the charge needed to switch from forward bias to reverse bias is dependent on the charge inserted into the BM during forward bias.

3.3.3 Rectification

The ability of a BM to rectify an ionic current, i.e. work as a diode and conduct better in the forward bias regime than the reverse bias regime, primarily originates from the difference in concentration in the junction between the two regimes. The current, and the rectification, through the BM is thus primarily dictated by the conductivity of the bulk of the IEMs in the forward bias regime and by the conductivity of the junction in the reverse bias regime. If co-ion transport in the BM is zero, the minimum conductivity of a BM in the reverse bias regime is dictated by the presence of protons and hydroxide ions due to self-ionization of water. This equilibrium reaction splits a water molecule into a proton and a hydroxide ion (H₂O −−*)−− H++ OH–) and gives 10 × 10−7_{m of protons and hydroxide ions at pH 7.}

3.3.4 Electric Field Enhanced Water Dissociation

A commonly observed phenomenon in BMs is that in the reverse bias regime, over a certain threshold voltage, the conductivity increases dramatically and non-linearly with voltage [27]. This gives a clear deviation from the low current reverse bias regime (Figure 3.8). At the same time pH changes are observed at the ends of the BM; the CEM side turns acidic and the 18

– + + + + + – – – – + + – –

+

–

(a) Forward bias

+ + – – + _– – + – – + + + + ++ + – – (b) Accumulation + + + – + – – –– + – ₊ – ₊ – ₊ + + (c) Ion crossing

Figure 3.9: (a) Ion motion in a BM during forward bias, and subsequent (b)

accumu-lation of ions and(c) co-ion crossing the BM due to high ion concentration.

+ + + – _– – + – – – + _Deplet ₊ ed

–

+

(a) Reverse bias

+ + – – + ₊ – – – + (b) Accumulated junction + + + + + – – – – – – – – + ++ (c) Ion extraction + + + – – – – + (d) Depleted junction

Figure 3.10: (a) Ion motion in a BM during reverse bias. If ions have (b) accumulated

in the junction, these ions must to be(c) extracted before the junction is (d) depleted.

– + H˜O OH– H+ H+ OH– H+ H+ OH– OH– OH– H+

–

+

Figure 3.11: EFE water dissocation. Water molecules in the BM juinction are split

AEM side alkali. This process, commonly called EFE water dissociation1 is attributed to an increased transport of H+ and OH– from the junction at a rate not explained by water self-ionization (Figure 3.11). EFE water dissociation has not only been observed for reversed biased BMs but also for IEMs, primarily anion-selective, above the Ilimdensity where an ion depleted zone is produced at the feeding side of the membrane [21].

The chemical reaction model

The mechanism for EFE water dissociation is not fully understood, but, as the name implies, the electric field is a main driving factor. In the chemical reaction model the increased water dissociation is attributed to a combination of the electric field and protonation-deprotonation reactions between catalytic species in the membrane and water molecules [24]. The local electric field across a depleted BM can be in the order of 10 × 108_{–10 × 10}9_{V/m, and} is thought to increase the forward reaction rate of water ionization [24]. Additionally, in the presence of a weak base (B) the following mechanism can occur [28]:

B + H₂O −−*)−− BH++ OH− BH++ H₂O −−*)−− B + H3O

+

where the net reaction produces one OH– and one H₃O+ (i.e. H+) from one water molecule. In the AEM of a BM, the weak base can be a tertiary amine. Similar mechanisms can be written for other types of catalytic groups, such as acids in a CEM or metal complexes, in the BM or for monopolar membranes.

Avoiding EFE water dissociation

EFE water dissociation is used in industrial applications to produce acids and bases [29], and it is therefore commercially interesting to develop membranes and catalysis groups for efficient EFE water dissociation [30, 31]. There are fewer reports on how to lower the EFE water dissociation. It has however been shown that AEMs composed of only quaternary amines and no tertiary amines does not produce the EFE water dissociation effect as long as the quaternary amines are not degraded to tertiary [32]. Non-amine based membranes, formed by immobilizing alkali ions to crown ethers, also show no EFE water dissociation [33]. A neutral layer inside the BM can also reduce the EFE water dissociation, as this increases the distance over which the potential drop occurs and thus lowers the electric field [34–36].

3.4 Electrodes in an Electrolyte

When an electrode is immersed into an electrolyte, electrons in the electrode and ions in the electrolyte interact. An electric potential difference between electrode and electrolyte will cause the formation of a polarized region along

1_{The term “water splitting” is also often used to describe this phenomena, but is easily}

confused for the electrochemical reaction that can occur at an electrode immersed in aqueous electrolyte.

– – – – – – – – – – + + + + + + – + + + – + + + + – + – + + + + – + Pot en tial Electrode Electrolyte Distance Helmholtz la yer Di˜usiv e layer

Figure 3.12: An EDL formed outside of a negatively charged electrode. The negative

charges on the electrode are compensated by cations in the Helmholtz layer, and an altered ion concentration profile in the diffusive layer.

the electrode-electrolyte interface called the electric double layer (EDL) [15]. The EDL consists of electric charges along the surface of the electrode compensated by an excess of oppositely charged ions outside of the electrode, i.e. a rearrangement of ions outside of the electrode (Figure 3.12).

In the Coüy-Chapman-Stern model, the electrolyte outside the electrode is divided into two layers: the Helmholtz layer and the diffuse layer. In the Helmholtz layer, the electronic charge on the surface of the electrode is compensated by oppositely charged ions that approach the electrode surface as close as possible and approximately at equal distance, forming the Helmholtz plane. The distance from the surface to the Helmholtz plane is dependent on the size of the ions, also taking in account their hydration shell. Due to the charge separation in the Helmholtz layer, which forms two oppositely charged planes with a small and well-defined separation, the potential drop across the Helmholtz layer is linear and steep.

The diffuse layer is located outside the Helmholtz layer and differs sig-nificantly from the Helmholtz layer in that it is composed of a gradient distribution of both anions and cations extending farther out into the elec-trolyte. In the diffuse layer, the electrode charge is compensated by an excess of oppositely charged ions and a shortage of same-charge ions (relative to the undisturbed bulk of the electrolyte). Compared to the Helmholtz layer, the potential drop in the diffuse layer extends farther out from the electrode and has en exponential decay.

By application of a potential between the electrode and the electrolyte (through a second electrode in the electrolyte), the charge of an EDL can increase, decrease, or change polarity. Charging of an EDL has a capacitive nature and the current associated with building up the EDL is thus called a capacitive, or non-Faradic, current. If no electrochemical active species are present in the electrolyte, or if the applied potential is lower than the

decomposition potential of the available species so that no electrochemical reaction can take place, no further current will flow through the systems once the EDL is fully formed. Such electrode is said to be polarizable within this potential range.

An EDL will also form if electrochemical active species, with decomposi-tion potential inside the applied potential, are present. The electrochemical reactions at the electrode, for example electrolysis of water, will give rise to an additional, Faradic, current. The Faradic current can be maintained even after the decay of the capacitive current. As the electrochemical reaction only occurs close to the electrode surface, and the reaction consumes available active species in this region, the Faradic current is often limited by diffusive transport of active species from the bulk electrolyte towards the electrode.

Chapter 4

Amyloids

Rudolph Virchow coined the term amyloid in 1854 to describe abnormal deposition found in tissues after staining with iodine [37]. The deposits were believed to contain starch, so the word amyloid, derived from the Greek word for starch, amylon, was used as the term for these structures. The content of amyloids were later discovered to not be starch but entirely represented by proteins. Amyloids are found in tissue samples from patients suffering from a variety of diseases. This lead to a classification of amyloid-related diseases based on the localization of the amyloid deposit [38]. Localized amyloidosis is organ or tissue specific, such as to the brain or heart. If the amyloid are not localized to a specific organ it is a systematic amyloidosis, i.e. amyloids are affecting a larger part of the body.

Further, specific proteins were found to generate the structural elements of the amyloids occurring in specific diseases. This lead to a second classi-fication of the amyloids, which is based on the assigned protein included. Numerous diseases causing dementia are coupled to proteins that forms amy-loid deposits in the brain [39], for example the amyamy-loid beta (Aβ) peptide in Alzheimer’s disease and the Alpha-synuclein peptide in Parkinson’s disease. Due to the projected burden on society for care taking of the increasing amount of people suffering of dementia and similar diseases as life expectancy increases [40], understanding, preventing, and curing amyloid related diseases is an important and active research topic today.

4.1 Amyloid Composition and Structure

The morphology of an amyloid can be stepwise divided into a macroscopically amorphous deposit (plaque), typically 1–100 µm in size, containing highly ordered nm-sized fibrils, that are composed of protein chains that have adopted a specific structure in the sub-nm domain (Figure 4.1).

Figure 4.1: Morphology of an amyloid. The plaque (surrounded by neuron cells)

R C C N O O H H H H

(a) Amino acid

R R R R H O H N C N C + H O H (b) Peptide bond

Figure 4.2: (a) Structure of a generic amino acid and (b) formation of a peptide

bond between two amino acids through a condensation reaction. R represens the amino acid side chain.

Proteins

Proteins are biological macromolecules present in all living cells and organisms. They serve a range of important functions such as catalyzing chemical reactions (enzymes), mediating intra- and extracellular signaling, or providing structural properties [41]. The building blocks of a protein are the amino acids (Figure 4.2a). These contain a carboxylic acid, an amine, and a side chain. The composition of the side chain defines the nature and function of the amino acids. Amino acids are joined through peptide bonding to form short (i.e. peptides) or long (i.e. proteins) polymer chains (Figure 4.2b).

Proteins are synthesized by cell ribosomes, using the genetic code of DNA, to yield well-defined sequences of amino acids in the chain. The sequence of the amino acids is the primary structure of the protein (Figure 4.3a). Local secondary structures, such as α helixes (Figure 4.3b), β sheets (Figure 4.3c), and turns, are formed through interaction between amino acids at different locations in the chain. The secondary structures, together with other non-local interactions between specific amino acids, define the tertiary structure, or fold, of the protein (Figure 4.3d). Multiple folded peptides can associate to form a quaternary structure. Some proteins do not adopt well-defined secondary or tertiary structures, but instead show a random-coil structure in their native state.

Structure of proteins in amyloids

In an amyloid, the folded or random-coiled protein has lost its native fold and/or function, and instead adopt a new, amyloid specific, fold [43]. This fold will vary by the nature of the protein and also by the microenvironment conditions. But, a key feature is a typical increase in β sheet or single β strand secondary structures (Figure 4.4). A second key feature of amyloids is the association of β sheets or β strands between proteins, yielding an extended β sheet structure perpendicular to the β strand direction. Such structures are called protofilament, and a bundle of protofilaments builds up the final fibril structure. Thus, the typical amyloid fibril appearance, with 24

Gly Pro Thr Gly Thr Gly

Glu Ser_Lys Cys Pro Leu Met Val Amino end Carboxyl end

(a) Primary (b) α helix (c) β sheet (d) Tertiary

Figure 4.3: (a) Primary, (b–c) secondary, and (d) tertiary structure of a protein.

Data for tertiary structure obtained from Kong et al. [42].

˜ sheet

Protoÿlament

Fibril

Figure 4.4: An extensive protein β sheets structure forms protofilaments, a subunit

in a amyloid fibril.

lengths in the range of µm and widths of tenths of nm, originates from long, often parallel, β sheets, where the individual β strands are short. Due to the rich β sheet structure the formed supramolecular assembly shows high stability.

4.2 Fibrillogenesis

Fibrillogenesis is the formation of fibers from its monomeric part, i.e. pro-teins, [44] and can be studied in vivo or, experimentally simpler, in vitro [45]. The specific molecular pathway to form amyloids differ from protein to protein and is not fully understood or known for all amyloidic proteins. A general pathway with common features can however be described, see Figure 4.5.

Nucleation growth

During in vitro studies of protein conversion into amyloid fibrils, a lag phase followed by a growth phase is typically observed [43]. In the lag phase no or little signs of fibrils are found, while during the growth phase fibrils are formed at an exponential rate (Figure 4.6). The cause of the lag phase is widely established to be the formation of a nucleus [43]. To form nucleus, a folded protein needs to partially unfold and adopt a new structure that has

Folded protein Unfolded peptide Partly folded intermediate Nucleus Preÿbrillar aggregate Mature ÿbril

Figure 4.5: Fibrillogenesis, formation of amyloid fibrils from folded or unfolded

protein chains, is a stepwise process.

some propensity to later form the β sheets. Similarly, a natively unfolded peptide needs to acquire such intermediate fold through partial folding. The rate of formation for the nucleus is typically slower than the elongation rate of the fibril [46], and this causes the lag phase. The lag phase can be shorten or removed by using conditions that increase the rate of nucleus formation or by adding preformed growth phase fibrils already containing nucleuses (“seeding”) [47].

Prefibrillar aggregates

In the lag phase oligomers and/or protofibrils1 _{are typically observed [48].} Oligomers are often nm-sized spherulites and may or may not bear resem-blance to the structure found in the mature fibrils. Protofibrils of Aβ on the other hand contain ∼ 20 monomers and show extensive β sheet structure. Protofibrils are highly cytotoxic, more so than mature fibrils, and might thus be the primary cause of the neurodegenerative damage observed in amyloid dementia diseases [49].

Mature fibrils

After nucleation, either by formation of oligomers or protofibrils, longer as-semblies are formed by elongation with monomeric or oligomeric proteins [50]. The growing amyloid fibers can also combine with each other to form bundles.

1_{Should not be confused with protofilament, which is a more “mature” structure.}

Time Fibr ils Lag phase Seeded Growth phase

Figure 4.6: Typical kinetics for amyloid aggregation, showing a lag phase followed

by a growth phase. By seeding the lag phase can be shortened.

The end result of this process is mature, insoluble fibers. In vivo, the fibers clump together to form plaques.

4.3 Factors for Amyloid Formation

Due to the wide diversity of amyloid forming proteins and the lack of correlation between primary and secondary structures for these proteins, it has been suggested that the ability to form amyloid aggregates is a general property of the polypeptide chain originating from its polymer nature, and not specific for certain amino acids or sequences [51, 52]. Nevertheless the formation of amyloids is (for most proteins) an abnormal function of the protein. Instead, certain factors can increase the likelihood or formation rate. Common factors important for the formation of amyloid aggregates, both in

vivo and in vitro, are protein concentration [53], protein stability [54], and

chemical (micro)environment [55].

In most inherent amyloid diseases mutations are found that destabilizes the protein and/or renders it more prone to form the initial amyloid structures. Several such mutations exist for Alzheimer’s disease [56], where for example mutations in the cleavage sites of the Aβ precursor peptide changes the ratio of Aβ(40) to Aβ(42) fragments produced. As Aβ(42) is more aggregation-prone than Aβ(40) these mutation, that gives an increase in Aβ(42), are more likely to cause Alzheimer’s disease.

Chemical factors known to affect the formation of amyloid aggregates are for example pH [57], ionic strength [51], and presence of specific ions such as Zn2+ [58] or Cu2+ [59]. Also, the presence of shear flow can induce aggregation [60].

4.4 Pathogenesis

The pathogenesis, i.e. the disease mechanism, for amyloid-related diseases shows significant variation depending on the protein and the aggregation location. Common themes can however be distinguished, primarily based on the localization of the aggregation, i.e. if the disease is systematic or local and, if local, to what organ.